Full disclosure, this opinion piece is completely written by me – a human – with no assistance from OpenAI’s ChatGPT or any other artificial intelligence large language model. Microsoft Word, ChatGPT’s distant cousin, corrected my spelling and reminded me to check my use of punctuation and some potentially ungrammatical clauses and sentences, but the thinking here is mine alone and I typed all of it.

ChatGPT could have easily completed a similar 1,000-word reflection in under five minutes. It might have produced the same kind of message with the right prompt and meticulous clarifying questions I supplied.

The University of Hong Kong last week announced a policy temporarily banning students from using ChatGPT and AI-based tools for coursework. Other Hong Kong universities, just like universities from the rest of the world, will have policies outlining what to do with content generated by AI and submitted to teachers in response to an assignment, exam or a full capstone requirement.

Polytechnic University, my home university, recently completed the first of many ChatGPT webinars attended by more than 500 participants, demonstrating the tool and discussing a range of teaching and assessment implications. PolyU has an AI and Data Analytics secondary major for undergraduates in several of its departments, so we are expected to be positioned to understand ChatGPT as much as possible and to be prepared.

In global academia, ChatGPT has been referred to as disruptive, upending our time-tested paradigms about education and learning. Its impact on the generation of knowledge, especially through writing, is a cause for major concern and even panic for many. One of the primary recipients of anxious thoughts and rushed policy declarations in schools is the essay, especially the high school and undergraduate kind.

Writing for The Atlantic, Stephen Marche argued that the undergraduate essay “has been the centre of humanistic pedagogy for generations. It is the way we teach children how to research, think and write. That entire tradition is about to be disrupted from the ground up”.

I see where these arguments are coming from, but this early in the game, I am cautiously swimming with lots of positive teaching and learning opportunities for myself and my students rather than drowning in apprehension.

I could be wrong, but I see ChatGPT and tools like it as transformative. These tools finally show AI’s march from machine learning to machine consciousness that could be leveraged and also controlled by teachers and students in the writing classroom and beyond.
Last November, I joined a group of applied linguists and English writing centre instructors from the southeastern United States in testing and exploring the deployment of ChatGPT in university English writing classrooms and what to do with it from assessment to policy. I was initially blown away primarily by how interactive and conversational it was, producing quick and extensive – if prompted – responses that went beyond culled or generic outputs.

You could argue with it, and it felt at times like you were communicating or debating with a polite expert or a very conscientious customer service representative who knew how to sell products well.

My English writing instructor colleagues immediately pointed out then that ChatGPT’s core weakness was that it couldn’t cite sources or use actual, accurate references. However, by late January, you could tell ChatGPT to include five sources in a 500-word essay about the prescience of Robert Frost’s poem Hyla Brook to global warming and add a bibliographic section using the Modern Language Association format.

The first output from my test prompt would likely receive a grade of B- to B following my typical rubric for an argumentative essay assignment. There was certainly room for improvement.

There also was a clear citation mistake, with one of Frost’s collections of poems identified as published in 2016. However, with follow-up strategic prompts and its polite ways of revising or disagreeing with me, it eventually got an A in my imaginary gradebook.

Accessing ChatGPT from Hong Kong via a virtual private network (VPN) and a US phone number is a challenge for me from time to time. Even so, I have enjoyed it, running out of academic essay prompts to test and focusing more on ephemeral topics on music, photography, dogs and all sorts of ideas that still generated output ranging from the interesting and creative content to the unoriginal or questionable.

I instructed ChatGPT to write a rhyming poem about how much I love and miss my dog Cocoa in the style of Mary Oliver. Although it was on target in parts, it read more like it was written by Bluey’s dad from the Disney animated television series than a Pulitzer Prize-winning poet. But ChatGPT writes poetry on demand, and it could still be poignant when one misses their dog.

If I stopped there, my Frost essay and Cocoa poem would certainly not contribute new knowledge. These activities didn’t make me think, and produce something I could claim as my own. I did not create or write what I had on my screen and shouldn’t be able to put my name on it, for doing so is hi-tech plagiarism.

But I see many entry points in a space like this for my students in the emerging AI-assisted writing classroom. I envision engagement opportunities that will prompt them to conduct their own additional research, read more and check for errors as they edit.

I will emphasise to my students that ethical considerations will always matter in my classroom and that I know what they know about the tool. Ultimately, I see my work as helping my students prepare for tomorrow because it’s here, now.